Unverified50% confidenceBenchmarkExact time
When solving complex programming problems, the number of internal reasoning tokens in the o1 model is often 10 to 50 times the final output.
1
Sources
50%
Confidence
Long-term
Relevance
7/24/2026
First Seen
Sources
Codex超高推理模式隐藏成本：一个任务耗尽70%额度的真相
redditr/OpenAI7/12/2026
Related Claims
Unverified根据文中量化估算，编程智能体任务中模型推理仅占总时间的 20%-30%，剩余 70%-80% 花在等待外部系统 I/O 响应72% similarUnverifiedReasoning-enhanced models such as the o1 series can cost several times to more than ten times the price of GPT-4o.67% similarUnverifiedIn High mode, the token consumption per chain-of-thought reasoning call may be 10 to 50 times that of Instant mode.67% similarUnverifiedModels with larger parameter counts and more reasoning steps may consume 10–50 times the GPU compute of lightweight models.66% similarUnverifiedBefore outputting a final answer, a reasoning model's internal chain of thought may consume 3,000 to 20,000 tokens, far exceeding the approximately 500 tokens of an ordinary model.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/608789API
curl https://kongchang.com/api/v1/knowledge/claims/608789MCP
get_claim(id=608789)